helping to satisfy the very high intrinsic energy demand of melting. As for the second, it transports a very hot (hence low‐viscosity) melt along the surface area, thereby facilitating the escape of rising bubbles. This effect is especially pronounced in the pattern shown in Figure 2a while in Figure 2b, a large portion of the melt does not reach the surface at all.
Table 4 Scheme for final batch adjustment with sodium sulphate set to 4 kg/t glass for a targeted redox number R of −24.
| Raw material i | mI(i)a | mII(i)b | mIII(i)c | R(i) | R(i)·mIII(i) |
|---|---|---|---|---|---|
| kg/t glass | kg/t glass | kg/2000 kg sand | |||
| Sand | 674.28 | 674.28 | 2000.00 | ||
| Feldspar | 44.02 | 44.02 | 130.57 | ||
| Calumite | 44.02 | 44.02 | 130.57 | −0.073 | −9.53 |
| Dolomite | 123.39 | 123.39 | 365.99 | ||
| Limestone | 70.61 | 70.61 | 209.44 | ||
| Soda ash | 231.99 | 229.01 | 679.26 | ||
| Sulfate | 4.00 | 11.86 | 0.67 | 7.95 | |
| Carbon | 1.13 | 3.35 | −6.70 | −22.46 | |
| Redox number R = ∑ R i ·m III (i) | −24.04 | ||||
∑ Ri·mIII(i) matches the target value R = −24.
a Batch composition as calculated in Table 2.
b Batch composition with 4 kg of sulfate added, soda ash reduced accordingly.
c Batch composition normalized to 2000 kg of sand; amount of carbon varied until the sum.
Figure 2 Convection cells (vortices) in the melting tank of a glass furnace (vertical projection): (a) float glass furnace (side port firing), transit to the refining zone indicated by the dotted vertical line on the right‐hand side; (b) end port fired container glass furnace, transit to the refiner through a narrow opening at the lower right (the “throat”); D = depth of the tank.
At a given pull rate p in t/h, the nominal overall dwell time of the melt in the furnace is
(1)
where ρ is the density of the melt and L·W·D the volume of the melting tank. Depending on the size of the furnace and the targeted glass quality (in terms of residual bubbles), τNOM ranges from 20 to 40 hours. During this time, the average volume element circles 2–6 times in vortex 1, and about twice in vortex 2. For a detailed analysis of the role of the flow pattern on melting and fining, see [3, 4]. The process of refining (M4) already starts at the descent of vortex 2; it is completed in a subsequent compartment termed refiner, which is thermally separated from the melter. Thermal separation is accomplished either with a vertical wall leaving an opening of about 0.5 × 0.3 m2 cross section (the throat) at half width right above the bottom of container‐glass furnaces (Chapter 1.5), or by an area of moderately narrowed width (the waist) in float glass furnaces (Chapter 1.4). For the sake of glass quality (i.e. homogeneity), it is mandatory to keep the position of the hot spot constant at any pull rate.
4.2 The Chemistry of Melting
The first high‐temperature process is primary melting of the batch. It is typically accomplished within one hour and is characterized by a very high energy demand and the release of CO2 from the carbonated raw materials. For a glass batch, the individual reactions involved are the following ones:
1 Physical melting of salt‐like raw materials, Na2CO3, Na2SO4, NaNO3, NaOH, NaCl, etc.
2 Evaporation of batch and hydrate water in the temperature range 100–600 °C.
3 Decomposition of limestone and dolomite:
4 Formation of a double carbonate: Na2CO3 + CaCO3 → Na2Ca(CO3)2 near 785 °C; in conventional batches, this is a side reaction of only minor importance.
5 Formation of silicate melts. These reactions assume noticeable turnover rates only after actual melting of soda ash:
Formation of ternary melts (eutectic NS–NS2–N2CS3: 821 °C, eutectic
1 Sulphate coal reaction: Na2SO4 + 4CO → Na2S + 4CO2 at approx. 900 °C.Figure 3 Liquidus temperatures and viscosities (dPa·s) of primary melts formed during the early stages of batch melting. (a) Binary salt‐like melts formed from soda ash and another compound. (b) Primary oxide melts in the systems Na2O–SiO2, Na2B4O7–SiO2, and B2O3–SiO2. Invariant points indicated by open circles.
2 Reactions